Notes of relevance to working with the Anaconda python distribution
- Full Anaconda
- Miniconda [my preference]
- Micromamba [minimalist shell install useful for headless machine install of a minimal set of packages]
- https://mamba.readthedocs.io/en/latest/installation/micromamba-installation.html
- Note that of course the installation url could change if you are automating installaions from the above
- For example into an AWS EC2 Deep Learning AMI
- https://mamba.readthedocs.io/en/latest/installation/micromamba-installation.html
- Mamba : the libmamba solver in the latest conda seems to be working well so far
-
Activation scripts are different between Linux and Windows
-
conda-pack is useful to transfer a working environment from one machine to another
-
It must be created on the same operating system as you are going to use it on as it is packaging necessary files/binaries
- Ubuntu -> Ubuntu
- Windows Server 2016 -> Windows Server 2016
-
There are other things you can do than an install to a default home directory
-
Need example
-
References
C:\Users\rscott>robocopy.exe C:\Users\rscott\Downloads\pangeo \\boringnetworkfolder\pangeo /E . /W:1 /R:1 /NFL /MT:64
-
echo Continuum/anaconda3/envs/*/lib/python*/site-packages/packageName | sed -E 's/[^ ]+envs\/([^/]+)\/lib[^ ]+/\1/g'
Continuum/anaconda3/condabin/conda.bat info --envs | awk '{print $1}' | xargs -ix Continuum/anaconda3/condabin/conda.bat list -n x | grep dask-geopandas
Continuum/anaconda3/condabin/conda.bat info --envs | awk '{print $1}' | xargs -ix Continuum/anaconda3/condabin/conda.bat list -n x | grep geoapps